Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/304745025

Imaginghydraulicfracturesbymicroseismicmigrationfordownholemonitoringsystem

ArticleinPhysicsofTheEarthandPlanetaryInteriors·June2016

DOI:10.1016/j.pepi.2016.06.010

CITATIONS

0

READS

45

2authors,including:

HaijiangZhang

UniversityofScienceandTechnologyofChina

94PUBLICATIONS1,558CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyHaijiangZhangon17October2016.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinblueareaddedtotheoriginaldocument

andarelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Physics of the Earth and Planetary Interiors

journal homepage: www.elsevier .com/locate /pepi

Imaging hydraulic fractures by microseismic migration for downholemonitoring system

http://dx.doi.org/10.1016/j.pepi.2016.06.0100031-9201/� 2016 Elsevier B.V. All rights reserved.

⇑ Corresponding author.E-mail address: zhang11@ustc.edu.cn (H. Zhang).

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fractures by microseismic migration for downhole monitoring system. PhyPlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Ye Lin, Haijiang Zhang ⇑Wantai Microseismic Lab of School of Earth and Space Sciences, University of Science and Technology of China, 96 Jinzhai Road, Hefei, Anhui 230026, China

a r t i c l e i n f o a b s t r a c t

Article history:Received 6 June 2016Accepted 21 June 2016Available online xxxx

Keywords:Hydraulic fracturingDownhole microseismic monitoringReverse time migrationStimulated reservoir volumeScattered waves

It has been a challenge to accurately characterize fracture zones created by hydraulic fracturing frommicroseismic event locations. This is because generally detected events are not complete due to the asso-ciated low signal to noise ratio and some fracturing stages may not produce microseismic events even iffractures are well developed. As a result, spatial distribution of microseismic events may not well repre-sent fractured zones by hydraulic fracturing. Here, we propose a new way to characterize the fracturedzones by reverse time migration (RTM) of microseismic waveforms from some events. This is based onthe fact that fractures filled with proppants and other fluids can act as strong scatterers for seismic waves.Therefore, for multi-stage hydraulic fracturing, recorded waveforms frommicroseismic events induced ina more recent stage may be scattered by fractured zones from previous stages. Through RTM of microseis-mic waveforms in the current stage, we can determine fractured zones created in previous stages byimaging area of strong scattering. We test the feasibility of this method using synthetic models with dif-ferent configurations of microseismic event locations and borehole sensor positions for a 2D downholemicroseismic monitoring system. Synthetic tests show that with a few events fractured zones can bedirectly imaged and thus the stimulated reservoir volume (SRV) can be estimated. Compared to the con-ventional location-based SRV estimation method, the proposed new method does not depend on thecompleteness of detected events and only a limited number of detected and located events are necessaryfor characterizing fracture distribution. For simplicity, the 2D model is used for illustrating the concept ofmicroseismic RTM for imaging the fracture zone but the method can be adapted to real cases in thefuture.

� 2016 Elsevier B.V. All rights reserved.

1. Introduction

Hydraulic fracturing is an engineering tool to efficiently recoveroil/gas from unconventional hydrocarbon reservoirs of low-permeability, including tight sand gas/oil and shale gas/oil. Forhydraulic fracturing, fractures are stimulated by pumping highpressure fluids into reservoirs through the treatment well. Duringthe fracking process, fractures are created by shear or tensile fail-ure generating observed induced microseismic events. Microseis-mic signals can be observed by surface or downhole monitoringgeophones. Compared to (near) surface geophones, downhole geo-phones installed in a borehole close to the fracking well are moreoften used for microseismic monitoring because microseismic sig-nals have higher signal to noise ratio (SNR). Microseismic eventsare distributed approximately around the face and tips of fractures,thus the microseismic cloud provides insight into fracture location,

height, length, orientation, and complexity (Maxwell et al., 2010).Generally, the main purpose of microseismic monitoring is toknow how fractures are created, where they are distributed, andthe characteristics of fractures. Using microseismic monitoring,fracking design can be optimized in order to improve the successratio of hydraulic fracturing and to avoid triggering large micro-seismic events along pre-existing faults (Maxwell, 2013).

For microseismic monitoring, the current efforts are generallyfocused on detecting microseismic events, determining their loca-tions and focal mechanisms. Hydraulic fracture distribution, aswell as stimulated reservoir volume (SRV) is generally determinedby microseismic locations (Urbancic et al., 1999; Mayerhofer et al.,2010). The more accurate the microseismic locations, the betterthe fracture distribution is characterized. Generally, two categoriesof microseismic location methodologies exist: (1) arrival-basedlocation methods and (2) migration-based location methods. Inthe case of high SNR (Daku et al., 2004), microseismic locationsare determined by picking first arrivals and then finding optimalsolutions by fitting absolute arrival times or arrival times

s. Earth

Fig. 1. 2D P-wave velocity model with a fracture zone at X = 1000 m. Thebackground velocity is 4200 m/s and the fracture zone is associated with a lowvelocity anomaly of 15% lower than the background velocity. Monitoring well islocated at X = 445 m with 36 sensors (denoted by triangles) ranging in depth fromZ = 975 m to Z = 2025 m with a spacing of 30 m. Microseismic events are (denotedby asterisks) located at (850 m, 2120 m), (850 m, 1975 m), (850 m, 2260 m),(862 m, 2048 m), (862 m, 2170 m), (830 m, 2107 m), (843 m, 2041 m), and (842 m,2235 m), respectively.

2 Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

differences. The other category is known as the migration-basedapproach by treating the ‘‘bright spot” as the seismic location afterstacking seismic energy on each node in the space domain thatseismic events may occur (Kao and Shan, 2004; Drew et al.,2005; Eisner et al., 2010; Chambers et al., 2010). For downholemicroseismic monitoring, back azimuth information of microseis-mic event needs to be used to constrain location in addition to firstarrivals. By combining microseismic locations with source mecha-nisms, it is helpful for interpreting fracture distribution and under-standing fracturing process (Gharti et al., 2011; Zhebel and Eisner,2014).

However, some microseismic events may be too weak to bedetected and for some reservoir layers hydraulic fracturing maynot necessarily induce microseismic events (Reshetnikov et al.,2010a,b; Das and Zoback, 2011; Zoback et al., 2012). As a result,the fracture network depicted solely by microseismic locations isbiased (Johri and Zoback, 2013; Sicking et al., 2012; Yang and

Fig. 2. (a) Snapshot of source wavefield at t = 0.1786 s. (b) Common-shot recordings for tsetup is shown in Fig. 1.

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Zoback, 2014). In fact, after hydraulic fracturing, some fracturesare filled with proppants and become strong scatterers for seismicwaves. Thus, for multi-stage fracturing, fractured areas in previousstages can act as scatterers for microseismic events induced byfracturing in later stages. As a result, coda waves of microseismicevents in later stages may contain information related to fracturedzones in previous stages. In this study, we will explore the possibil-ity of using these scattered waves to characterize the fracturedzones.

It is known that seismic migration is the most significant pro-cess in seismic exploration because of its ability to image the sub-surface structure using various kinds of wavefield. Prestack depthmigration has developed from Kirchhoff migration and beammigration to one-way wave-equation migration and reverse timemigration (Schneider, 1978; Hill, 1990; Claerbout, 1971; Baysalet al., 1983). In comparison, RTM is more suitable for imaging com-plicated structures because of its ability to handle various kinds ofwavefield including reflected, refracted, diffracted, turning andmultiple waves.

In recent years, the concept of seismic migration developed inseismic exploration with active sources has already been appliedto passive earthquake sources to image structure discontinuitiesin the earth at various scales. Chavarria et al. (2003) used bothscattered P- and S-waves of recorded waveforms from earthquakesrecorded in the pilot well of the San Andreas Fault Observatory atDepth (SAFOD) to image the fault zone structure through Kirchhoffmigration. Zhang et al. (2009a) used scattered P-waves of earth-quake waveforms to image reflectors around the SAFOD sitethrough a generalized Radon transform. Reshetnikov et al.(2010a) imaged the San Andreas Fault structure by applying theFresnel volume migration method to earthquake waveformsrecorded by downhole sensors. Reshetnikov et al. (2010b) alsoapplied the Fresnel volume migration method to induced high-frequency microseismic waveforms from the German ContinentalDeep Drilling program to obtain the fault structure.

In this study, we will test the ability of using microseismicevents to image fractured zones by the RTMmethod from scatteredwaves in microseismic recordings. For simplicity, we assume thefractures and microseismic events are located in the same two-dimensional plane as the downhole sensors, but the method canbe adapted to a more generalized three-dimensional case. We testthe feasibility of this method using different configurations ofmicroseismic event locations and sensor locations as well as differ-ent synthetic models.

he event located at (850 m, 1975 m) and the sample interval is 0.595 ms. The model

s by microseismic migration for downhole monitoring system. Phys. Earth

Fig. 3. RTM imaging result from 8 individual microseismic events (denoted by asterisks) for the model setup in Fig. 1.

Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx 3

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fractures by microseismic migration for downhole monitoring system. Phys. EarthPlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Fig. 4. Normalized and stacked RTM imaging result from 8 microseismic eventsshown in Fig. 3.

Fig. 5. Comparison of RTM imaging results for sensors located at different depthsfor the model setup shown in Fig. 1. (a) From Z = 1575 m to 2625 m. (b) FromZ = 775 m to 1825 m.

4 Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

2. Microseismic reverse time migration method

For downhole microseismic monitoring, a string of seismic sen-sors is installed in a borehole for monitoring microseismic events.In this case, the configuration of seismic sensors and sources issimilar to the vertical seismic profile (VSP) system except thesources are located around reservoirs for microseismic monitoringwhile they are located on surface for VSP. For this reason, we canadapt the VSP RTM method (Sun et al., 2011; Xiao and Leaney,2010) to downhole microseismic monitoring. The conventionalreverse time migration method consists of four generalized steps:(1) stimulate the source and forward propagate the source wave-field, (2) set the microseismic recordings as boundary conditionsand backward propagate the wavefield, (3) apply the imaging con-dition to the forward-propagating wavefield and the backward-propagating wavefield and get the imaging result for this micro-seismic event, and (4) sum over imaging results from all eventsto get the final migration image.

Here, we formulate the RTM method based on a 2D acousticwave equation (Zhang and Sun, 2008; Sun et al., 2011). For theborehole array, we need to compute forward-propagating wave-field pF with event originating at ðxs; zxÞ by

1c2

@2

@t2�r2

!pFðx; z; tÞ ¼ dððx; zÞ � ðxs; zsÞÞf ðtÞ ð1Þ

In the above equation, c ¼ cðx; zÞ is the velocity model, f ðtÞ isthe microseismic source signature, and r2 is Laplace operator.

For the common-event recordings Dðxr ; z; xs; zs; tÞ by boreholearray at (xr ; z), we need to back propagate wavefield pB based onthe following equation:

1c2

@2

@t2�r2

� �pBðx; z; tÞ ¼ 0

pBðx ¼ xr ; z; tÞ ¼ Dðxr; z; xs; zs; tÞ

(ð2Þ

To simulate both the source wavefield and the receiver wave-field, we adopted the finite difference method (FDM) to solve theabove equations with perfectly matched layer (PML) boundarycondition.

The imaging condition for the conventional RTM is zero-lagcross-correlation of the source wavefield and the receiver wave-field. The images are given by

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Iðx; zÞ ¼Z t

0Sðx; z; sÞRðx; z; t � sÞds ð3Þ

where t is time, S and R denote the forward-propagating sourcewavefield and backward-propagating receiver wavefield at spatialcoordinate ðx; zÞ, and Iðx; zÞ represents the imaging result at ðx; zÞ.The final migration image is obtained by stacking the images foreach event.

The major drawback of conventional RTM is the low-frequencynoise caused by cross-correlation of unwanted wavefield, whichadversely affects the imaging resolution. Thus, the suppression oflow-frequency noise is an active research area in RTM and manymethods have been proposed (Baysal et al., 1984; Loewenthalet al., 1987; Guitton et al., 2006; Zhang and Sun, 2008; Yoon andMarfurt, 2006). In this study, we adopted the wavefield-separation method (Liu et al., 2007, 2011; Xiao and Leaney,2010), which applies the cross-correlation imaging condition tosource wavefield and receiver wavefield at different directions, asfollows,

Iðx; zÞ ¼Z t

0S�ðx; z; sÞR�ðx; z; t � sÞds ð4Þ

s by microseismic migration for downhole monitoring system. Phys. Earth

Fig. 6. Comparison of imaging results using (left) the individual event located at (850 m, 2120 m) and (right) all 8 events together in different cases of having (top) 12 sensors,(middle) 24 sensors and (bottom) 36 sensors. The span of these sensors is the same in depth but with different spacing of 90, 45 and 30 m from top to bottom. Other symbolsare the same as those in Fig. 3.

Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx 5

Here, the superscripts ‘‘+” and ‘‘�” denote the different direc-tions along the horizontal or vertical directions.

3. Synthetic test of downhole microseismic migration method

For simplicity, we assume that microseismic sources, downholesensors and fractures are located in a 2D plane and accordinglydesign a 2D acoustic model for testing the microseismic migrationmethod introduced in the above section. Furthermore, we assumethat both source locations and velocity model are known in thenumerical testing of the method.

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Fig. 1 shows the configuration of microseismic events, down-hole sensors and an oval-shaped fracture. The fracture is locatedat X = 1000 m and is assumed to be induced by a previoushydraulic fracturing stage. Thirty-six sensors are installed in themonitoring well at X = 445 m. The sensors range in depth fromZ = 975 m to Z = 2025 m with a spacing of 30 m. There are 8microseismic events, which are assumed to be induced by a laterfracking stage between the fracture and the monitoring well.These microseismic events are randomly distributed in the frac-turing zone.

We used a 40 Hz Ricker wavelet as the source to model syn-thetic waveforms based on the seismic acoustic equation. The

s by microseismic migration for downhole monitoring system. Phys. Earth

Fig. 7. Comparison of the RTM images when the background velocity models are complex. (a) The background velocity model contains several layers with different velocitiescompared to that shown in Fig. 1; (b) normalized and stacked RTM image from 8 microseismic events for the model setup in (a); (c) the same as (a) but with a normal fault tothe east of the fracture zone; (d) normalized and stacked RTM image for the model setup in (c). The model layers are denoted by dashed lines in (b) and (d).

6 Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

common-event recordings are 2 min long. A 10th-order accuracycentral difference stencil is applied to solve the acoustic equationwith the perfectly matched layer (PML) boundary condition. Fromthe snapshot of forward-propagating source wavefield att = 0.1786 s (Fig. 2a) and the recordings (Fig. 2b), the scatteredwaves from the fractured zone can be clearly seen. For microseis-mic migration, the background velocity model is used. The respec-tive RTM imaging results for 8 individual microseismic events areshown in Fig. 3. All images clearly reveal that there exists an impe-dance anomaly around X = 1000 m where the oval-shaped fractureis located. However, the images are different for different events,indicating that the contribution to imaging is different for eachevent because of their different locations relative to the fracturedzone. By stacking the normalized RTM imaging result from eachindividual event, the fractured zone is better imaged due to thebetter data coverage (Fig. 4). The image of the fracture shapebecomes clearer and the upper tip of the fracture is clearly imaged.In comparison, the lower part of the fracture is not imaged as wellas the upper part. This is because there are observations from only8 events distributed in a limited aperture and the sensors arelocated relatively higher than the fracture. As a result, sensors inthe monitoring well cannot receive enough wavefield that containsinformation related to the lower part as well as the bottom tip ofthe fracture.

There are clear artifacts in the RTM imaging result for eachevent due to the limited spatial coverage of wavefield by using onlyone event (Fig. 3), and the artifacts can be suppressed by using 8

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

events (Fig. 4). There are also some artifacts appearing to the westof the monitoring well (Figs. 3 and 4). This kind of artifacts has alsobeen noted when applying the RTMmethod to vertical seismic pro-file system, which is mostly caused by the limited spatial coverageof the observation system and can be alleviated by the increase ofsensor spatial coverage (Sun et al., 2011).

To further test how the sensor spatial coverage affects thedownhole microseismic RTM imaging result, we shift the sensorsboth downward and upward. For the sensors located fromZ = 1575 m to Z = 2625 m, the summed RTM image becomes moreaccurate and the shape of the fracture zone becomes clearerbecause more wavefields coming from the lower part of fracturecan be received (Fig. 5a). If the sensors are shifted upward to belocated in the range of Z = 775 m to Z = 1825 m, the fracture ismore poorly resolved especially for the lower part of the fracture(Fig. 5b). This indicates that with a larger aperture of sensors, thefracture can be better characterized because wavefield from differ-ent directions can be received.

We also test how the spacing interval of borehole sensors affectthe downhole microseismic RTM imaging result. For comparison,we used 12 and 24 borehole sensors from Z = 975 m to 2025 mwith an interval of 90 m and 45 m, respectively. The fracture isclearly imaged from using one event or all 8 events with 12, 24or 36 sensors (Fig. 6). It is clear, however, that the fracture isslightly better imaged with more sensors. For the case of using only12 sensors, arc-shaped spatial aliasing can adversely affect theimaging resolution due to the low spatial sampling of seismic

s by microseismic migration for downhole monitoring system. Phys. Earth

Fig. 8. The RTM images from microseismic recordings with different levels of noise. (a) The recordings from a microseismic event by adding Gaussian distributed noise withSNR of 2. (b) Normalized and stacked RTM image for 8 events for the noise level in (a). (c) The recordings from a microseismic event by adding Gaussian distributed noise withSNR of 0.5. (d) Normalized and stacked RTM image for 8 events for the noise level in (c). The model setup is the same as Fig. 1.

Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx 7

wavefield. This problem can be mitigated by seismic data recon-struction, such as the compressive sensing method (Tang, 2010).

In reality, the background model is not homogeneous and forthe case of shale gas development the model can be approximatedby a layer model and the hydraulic fracturing is conducted in thetarget shale layer. To assess how the subsurface structures affectthe fracture imaging, we test two models: one by adding severallayers to the homogeneous background (Fig. 7a), and the otherby adding a normal fault in addition to layers to the homogeneousbackground (Fig. 7c). Except for the background velocity model, allthe other model setup parameters are the same as those in Fig. 1.For microseismic migration in this case, we again assume the back-ground velocity model is known. The tests using two models showthat even with more complex structures in the background model,the fracture zone can still be clearly imaged (Fig. 7b, d). At the sametime the layer interfaces below the microseismic events are alsoimaged, especially for the first layer interface (Fig. 7b, d). Further-more, the steep normal fault to the east of the fracture zone is alsoclearly imaged (Fig. 7d). This indicates that for the real case ofhydraulic fracturing, it is possible to image both fracture zoneand the shale layers by using microseismic events.

In practice, microseismic signals may be contaminated by noise.To test how noise affects the imaging result, we add Gaussian dis-tributed noise to the recordings with signal-to-noise ratio (SNR) of2 (Fig. 8a) and 0.5 (Fig. 8c), respectively. The normalized andstacked microseismic RTM images with all 8 microseismic events

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

show that even with noise the fracture zone can still be clearlyimaged although migration images are contaminated to somedegree (Fig. 8b, d). This demonstrates that the wavefield-separation imaging condition method used in microseismic RTMbehaves well in suppressing the imaging noise.

It is inevitable that the microseismic event is generally associ-ated with some uncertainty in its location and origin time. How-ever, with more advanced microseismic location methods such asthe double-difference location method (Waldhauser andEllsworth, 2000), the microseismic event locations are very accu-rate and the associated uncertainties could be on the order of10 m (Castellanos and Van der Baan, 2013; Li et al., 2013). Here,we further test how location errors affect the RTM imaging resultsby perturbing event locations randomly within 10 meters (Fig. 9a),20 m (Fig. 9c) and 50 m (Fig. 9e), respectively. Fig. 9b, Fig. 9d andFig. 9f show the normalized and stacked RTM images correspond-ing to perturbed event locations in Fig. 9a, Fig. 9c and Fig. 9e,respectively. Compared with the image using true event locationsshown in Fig. 4, the RTM image is almost not affected by randomlyperturbing locations within 10 m (Fig. 9b). In comparison, greaterlocation errors may adversely affect the RTM result and it can beseen that the fracture structures are slightly distorted by randomlyperturbing locations within 20 m (Fig. 9d). If we further perturbevent locations randomly within 50 m on the order of the halfwavelength (Fig. 9e), the fracture zone is more distorted but it isstill clearly recognizable in the migrated image (Fig. 9f).

s by microseismic migration for downhole monitoring system. Phys. Earth

Fig. 9. Normalized and stacked RTM images (right) from 8 microseismic events when their locations are randomly perturbed at different degrees (left). (a) and (b) showlocations randomly perturbed within 10 m and the corresponding RTM image. (c) and (d) show locations randomly perturbed within 20 m and the corresponding RTM image.(e) and (f) show locations randomly perturbed within 50 m and the corresponding RTM image. White asterisks denote true seismic locations and red asterisks denoteperturbed locations.

8 Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

4. Discussion and conclusions

In this study, we propose to take advantage of microseismicscattered waves and apply the RTM method to directly image thefracture zone by hydraulic fracturing. Unlike the conventionallocation-based fracture characterization method that uses micro-seismic locations as a proxy for determining the fracture distribu-tion, the microseismic migration method that we propose to use inthis study does not need to detect and use all microseismic eventsto accurately characterize the fracture zone. As long as a few

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

microseismic events with high SNR are detected and located, theycan be used to characterize the fractured zone. The numerical testsshowed that with only 8 microseismic events, we are able toclearly characterize the fracture zone.

For microseismic monitoring, it is important to estimate theSRV. More importantly, it is better to estimate the effective SRVwhere the proppants exist and fractures are conductively con-nected (Mayerhofer et al., 2010; Johri and Zoback, August 2013).However, for the zone where the microseismic events are detectedand located, it does not necessarily indicate the existence of

s by microseismic migration for downhole monitoring system. Phys. Earth

Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx 9

proppants and opened fractures there. Therefore, it has been achallenge to use microseismic monitoring to characterize whereproppants migrate. With the microseismic migration method, itis possible to differentiate the region with or without proppantsand opened fractures. Even if the region is associated with micro-seismic events, if it does not show strong scattering effect the frac-tures may be closed due to the lack of support by proppants.

We showed that for different sensor positions relative to thefracture zone, the RTM imaging result is different. The monitoringsystem having better spatial sensor coverage can better image thefracture zone. Therefore, if it is possible, the sensors should beinstalled around the same depth zone of fracking. It is also notedthat different sensor spacing intervals also affect the imaging resultand for the case of large spacing interval the spatial aliasing causedby low spatial sampling of the wavefield will impart the imagingresolution. In addition to imaging the fracture zone, from differentsynthetic tests it is also noted that the backward-propagatingwavefield is focused at the source locations (Fig. 3).

In real cases, the microseismic signals are contaminated withnoise to some degrees. The synthetic tests by adding differentlevels of noise to microseismic recordings show that thewavefield-separation imaging condition method used in microseis-mic RTM can satisfactorily deal with the noisy data to recover thefractured zone. We also test how microseismic event locationerrors affect the RTM imaging results and it shows that the fracturezone can be biased to some degree but can still be clearly imagedeven when the event locations are randomly perturbed within50 m, approximately on the order of half wavelength. With moreadvanced microseismic location methods available, the locationerrors could well be on the order of 10 s of meters. The uncertaintyin the origin time bears a similar effect on the RTM imaging resultas location errors. It is also possible that different microseismicevents may have different source time functions. However, dueto the nature of high frequency and limited bandwidth for micro-seismic signals, the corresponding source time functions can bewell approximated by a smooth ramp function with a very shortduration (Li et al., 2011a,b). In addition, most of microseismicevents from the same fracking stage generally have similar focalmechanisms. Therefore, it is expected that source signatures ofmicroseismic events would not greatly affect the RTM image.

In these tests, we assume the background velocity model usedfor microseismic migration is known. In real cases, the velocitymodel can be well estimated from sonic well logging and cali-brated by perforation shots (Zhang et al., 2016). Furthermore, itcan also be estimated by microseismic velocity tomography usingarrival times frommicroseismic events (Zhang et al., 2009b). In thisstudy, we only use the reverse time migration method to the scat-tered P waves, but the method can also be adapted to use scatteredS waves based on the elastic wave equation. We emphasize thatthe goal of this study is to show the concept of microseismicRTM to image the fractures. For the application of this method tothe real microseismic data, we need to adapt the microseismicRTM to the 3D case. One way to do this is to use the 3D acousticequation and modify the Eqs. (1) and (2) accordingly to realizethe 3D microseismic RTM. Another way is to apply the 2D RTMto each microseismic event that forms a 2D plane with the verticalmonitoring array. By stacking RTM images from different micro-seismic events, we may be able to obtain the 3D RTM image. Tobetter take care of uncertainties in the origin time, location andsource signature of microseismic event, we could apply the elasticreverse time migration imaging with both microseismic scatteredP and S waves (Xiao and Leaney, 2010). When combined withmicroseismic locations and focal mechanisms, microseismicmigration can help us to better understand the fracture network,and thus to better optimize the fracturing design and to enhancethe oil/gas recovery.

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

Acknowledgements

This research is supported by Natural Science Foundation ofChina under the grant No. 41274055. We are grateful for construc-tive comments from two anonymous reviewers that are helpful forimproving this paper.

References

Baysal, E., Kosloff, D.D., Sherwood, J.W.C., 1983. Reverse time migration. Geophysics48 (11), 1514–1524. http://dx.doi.org/10.1190/1.1441434.

Baysal, E., Kosloff, D.D., Sherwood, J.W.C., 1984. A two-way non reflecting waveequation. Geophysics 49, 132–141. http://dx.doi.org/10.1190/1.1441644.

Castellanos, F., Van der Baan, M., 2013. Microseismic event locations using thedouble-difference algorithm. CSEG Recorder 38 (3), 26–37.

Chambers, K., Kendall, J., Brandsberg-Dahl, S., et al., 2010. Testing the ability ofsurface arrays to monitor microseismic activity. Geophys. Prospect. 58 (5), 821–830. http://dx.doi.org/10.1111/j.1365-2478.2010.00893.x.

Chavarria, J.A., Malin, P., Catchings, R.D., et al., 2003. A look inside the San Andreasfault at Parkfield through vertical seismic profiling. Science 302 (5651), 1746–1748. http://dx.doi.org/10.1126/science.1090711.

Claerbout, J.F., 1971. Toward a unified theory of reflector mapping. Geophysics 36(3), 467–481. http://dx.doi.org/10.1190/1.1440185.

Daku, B.L.F., Salt, J.E., Sha, L., 2004. An algorithm for locating microseismic events,Canadian Conference on Electrical and Computer Engineering. IEEE, pp. 2311–2314, <http://dx.doi.org/10.1109/CCECE.2004.1347708>.

Das, I., Zoback, M.D., 2011. Long-period, long-duration seismic events duringhydraulic fracture stimulation of a shale gas reservoir. SEG Tech. ProgramExpanded Abstr., 1473–1477 http://dx.doi.org/10.1190/1.3627480.

Drew, J., Leslie, H., Armstrong, P., et al., 2005. Automated microseismic eventdetection and location by continuous spatial mapping, SPE Annual TechnicalConference and Exhibition. http://dx.doi.org/10.2118/95513-MS.

Eisner, L., Hulsey, B.J., Duncan, P., et al., 2010. Comparison of surface and boreholelocations of induced seismicity. Geophys. Prospect. 58 (5), 809–820. http://dx.doi.org/10.1111/j.1365-2478.2010.00867.x.

Gharti, H.N., Oye, V., Kühn, D., et al., 2011. Simultaneous microearthquake locationand moment-tensor estimation using time-reversal imaging, SEG AnnualMeeting.

Guitton, A., Kaelin, B., Biondi, B., 2006. Least-square attenuation of reverse timemigration artifacts. SEG Tech. Program Expanded Abstr., 2348–2352 http://dx.doi.org/10.1190/1.2370005.

Hill, N.R., 1990. Gaussian beam migration. Geophysics 55 (11), 1416–1428. http://dx.doi.org/10.1190/1.1442788.

Johri, M., Zoback, M.D., 2013. The evolution of stimulated reservoir volume duringhydraulic stimulation of shale gas formations, Unconventional ResourcesTechnology Conference, Denver, Colorado, 12-14 August 2013, pp. 1661–1671. http://dx.doi.org/10.1190/urtec2013-170.

Kao, H., Shan, S.J., 2004. The source-scanning algorithm: mapping the distribution ofseismic sources in time and space. Geophys. J. Int. 157 (2), 589–594. http://dx.doi.org/10.1111/j.1365-246X.2004.02276.x.

Li, J., Sadi, Kuleli H., Zhang, H., et al., 2011a. Focal mechanism determination ofinduced microearthquakes in an oil field using full waveforms from shallow anddeep seismic networks. Geophysics 76 (6), WC87–WC101. http://dx.doi.org/10.1190/geo2011-0030.1.

Li, J., Zhang, H., Kuleli, H.S., et al., 2011b. Focal mechanism determination usinghigh-frequency waveform matching and its application to small magnitudeinduced earthquakes. Geophys. J. Int. 184 (3), 1261–1274. http://dx.doi.org/10.1111/j.1365-246X.2010.04903.x.

Li, J., Zhang, H., Rodi, W.L., et al., 2013. Joint microseismic location and anisotropictomography using differential arrival times and differential backazimuths.Geophys. J. Int. 195 (3), 1917–1931. http://dx.doi.org/10.1093/gji/ggt358.

Liu, F., Zhang, G., Morton, S., et al., 2007. Reverse-time migration using one-waywavefield imaging condition. In: 77th Annual International Meeting, SEG, pp.2170–2174. http://dx.doi.org/10.1190/1.2792917.

Liu, F., Zhang, G., Morton, S., et al., 2011. An effective imaging condition for reverse-time migration using wavefield decomposition. Geophysics 76, s29–s39. http://dx.doi.org/10.1190/1.3533914.

Loewenthal, D., Stoffa, P.A., Faria, E.L., 1987. Suppressing the unwanted reflectionsof the full wave equation. Geophysics 52, 1007–1012. http://dx.doi.org/10.1190/1.1442352.

Maxwell, S.C., Rutledge, J., Jones, R., et al., 2010. Petroleum reservoircharacterization using downhole microseismic monitoring. Geophysics 75 (5).http://dx.doi.org/10.1190/1.3477966, 75A129–75A137.

Maxwell, S., 2013. Unintentional seismicity induced by hydraulic fracturing. Can.Soc. Exp. Geophys. Rec. 38 (8), 40–49.

Mayerhofer, M., Lolon, E., Warpinski, N., et al., 2010. What is stimulated reservoirvolume? SPE Prod. Oper. 25 (1), 89–98. http://dx.doi.org/10.2118/119890-PA.

Reshetnikov, A., Buske, S., Shapiro, S.A., 2010a. Seismic imaging using microseismicevents: Results from the San Andreas Fault System at SAFOD. In: J. Geophys.Res. 115 (B12). http://dx.doi.org/10.1029/2009JB007049.

Reshetnikov, A., Kummerow, J., Buske, S., 2010b. Microseismic imaging from asingle geophone, KTB[C]//2010 SEG Annual Meeting. Society of ExplorationGeophysicists. http://dx.doi.org/10.1190/1.3513252.

s by microseismic migration for downhole monitoring system. Phys. Earth

10 Y. Lin, H. Zhang / Physics of the Earth and Planetary Interiors xxx (2016) xxx–xxx

Schneider, W.A., 1978. Integral formulation for migration in two and threedimensions. Geophysics 43 (1), 49–76. http://dx.doi.org/10.1190/1.1440828.

Sicking, C., Vermilye, J., Geiser, P., et al., 2012. Permeability field imaging frommicroseismic. SEG Tech. Program Expanded Abstr., 1–5 http://dx.doi.org/10.1190/segam2012-1383.1.

Sun, W.B., Sun, Z.D., Zhu, X.H., 2011. Amplitude preserved VSP reverse timemigration for angle-domain CIGs extraction. Appl. Geophys. 8 (2), 141–149.http://dx.doi.org/10.1007/s11770-011-0277-1.

Tang, G., 2010. Seismic Data Reconstruction and Denoising Based on CompressiveSensing and Sparse Representation. Tsinghua University, Beijing.

Urbancic, T.I.J., Shumila, V.J., Rutledge, J.T.J., 1999. Determining hydraulic fracturebehavior using microseismicity. Vail Rocks 1999. In: The 37th US Symposiumon Rock Mechanics (USRMS). American Rock Mechanics Association.

Waldhauser, F., Ellsworth, W.L., 2000. A double-difference earthquake locationalgorithm: method and application to the northern Hayward fault, California.Bull. Seismol. Soc. Am. 90 (6), 1353–1368. http://dx.doi.org/10.1785/0120000006.

Xiao, X., Leaney, W.S., 2010. Local vertical seismic profiling (VSP) elastic reverse-time migration and migration resolution: salt-flank imaging with transmittedP-to-S waves. Geophysics 75 (2), S35–S49. http://dx.doi.org/10.1190/1.3309460.

Yang, Y., Zoback, M.D., 2014. The role of preexisting fractures and faults duringmultistage hydraulic fracturing in the Bakken Formation. Interpretation 2 (3),G25–G39. http://dx.doi.org/10.1190/INT-2013-0158.1.

Please cite this article in press as: Lin, Y., Zhang, H. Imaging hydraulic fracturePlanet. In. (2016), http://dx.doi.org/10.1016/j.pepi.2016.06.010

View publication statsView publication stats

Yoon, K., Marfurt, K.J., 2006. Reverse-time migration using the pointing vector.Explor. Geophys. 59, 102–107. http://dx.doi.org/10.1071/EG06102.

Zhang, H., Wang, P., van der Hilst, R.D., et al., 2009a. Three-dimensional passiveseismic waveform imaging around the SAFOD site, California, using thegeneralized Radon transform. Geophys. Res. Lett. 36 (23). http://dx.doi.org/10.1029/2009gl040372.

Zhang, H., Sarkar, S., Toksoz, M.N., Kuleli, H.S., Al-Kindy, F., 2009b. Passive seismictomography using induced seismicity at a petroleum field in Oman. Geophysics74 (6), WCB67. http://dx.doi.org/10.1190/1.3253059.

Zhang, J., Zhang, H., Zhang, Y., Liu, Q., 2016. Fast one-dimensional velocity modeldetermination for microseismic monitoring using station-pair differentialarrival times based on the differential evolution method. Phys. Earth Planet.Inter. http://dx.doi.org/10.1016/j.pepi.2016.06.003.

Zhang, Y., Sun, J., 2008. Practical issues of reverse time migration-true-amplitudegathers, noise removal and harmonic-source encoding, 70th EAGE Conference &Exhibition. http://dx.doi.org/10.3997/2214-4609.20147708.

Zhebel, O., Eisner, L., 2014. Simultaneous microseismic event localization andsource mechanism determination. Geophysics 80 (1), KS1–KS9. http://dx.doi.org/10.1190/segam2012-1033.1.

Zoback, M.D., Kohli, A., Das, I., 2012. The importance of slow slip on faults duringhydraulic fracturing stimulation of shale gas reservoirs, SPE AmericasUnconventional Resources Conference. Society of Petroleum Engineers. http://dx.doi.org/10.2118/155476-MS.

s by microseismic migration for downhole monitoring system. Phys. Earth